JP7326400B2 - Content processing method and content processing program - Google Patents

Description

The present invention relates to a content processing method and a content processing program.

Efficient search technology for content, including technical documents, is required to promote research and development, intellectual property strategies, and marketing strategies for company products and services in companies, research institutes, and the like. In order to prevent infringement of patent rights, acquire rights, and understand other companies' technologies, it is becoming important to obtain information efficiently without omissions, especially in searching patent documents.
As conventional techniques, there are the following techniques.

For example, for each of a plurality of evaluation target documents containing a plurality of words, an evaluation is received from the user as to whether it is a positive evaluation document related to the target theme or a negative evaluation document not related to the target theme, and each evaluation is performed. Words are extracted from the target document and classified into positive words appearing only in positive evaluation documents, negative words appearing only in negative evaluation documents, and common words appearing in both positive evaluation documents and negative evaluation documents. is extracted, and based on the appearance frequency of each word and the adjacency relationship with other words, the degree of theme relevance of the common word to the target theme is calculated (see, for example, Patent Document 1).

Also, unread information is input, and a set of information data and a teacher signal indicating whether one or more pieces of information consisting of one or more keywords are necessary or not is prepared in advance as teacher data, and is newly input. One or more keywords attached to the unread information, and pairs of the keywords and the teacher signals. If there are many pairs of teacher signals required for the keywords attached to the unread information, a large value is set to the number of unnecessary teacher signals. There is a technique for obtaining a need signal that predicts the user's need for unread information with a small value if the number of sets is large (see, for example, Patent Document 2).

Japanese Patent No. 5424393 Japanese Patent No. 3736564

The disclosed technology assists the operator to more efficiently grasp each content included in the set of contents when the operator is provided with a set of contents obtained by searching for contents including text. By doing so, the purpose is to reduce the labor of the operator.

The disclosed technique is a content processing method for determining the priority of presentation of each of a plurality of contents, comprising: identifying the plurality of contents; a plurality of keywords specified by an operator; receiving keyword information, including a weight corresponding to each; and calculating, for each of the plurality of contents, the product of the number of occurrences of each of the plurality of keywords and the weight corresponding to the keyword for each of the plurality of keywords. deriving the sum corresponding to each of the plurality of pieces of content by adding together to obtain a sum for each of the plurality of pieces of content; determining a degree of content processing.

Further, identifying the plurality of contents in the technology disclosed herein includes extracting a plurality of words related to the plurality of contents, and extracting the plurality of words from the plurality of words based on the plurality of words. presenting the plurality of words to an operator to identify keywords.

In addition, zero may be included in the possible values of the weights in the technology disclosed herein.

Further, the extracting in the technology disclosed herein may include extracting the plurality of words from predetermined portions of the plurality of contents.

In addition, the content in technology disclosed herein may include at least one of text, images, and audio.

Also, extracting the plurality of words in the technology disclosed herein includes receiving a positive or negative evaluation value given by the operator to each of the plurality of contents peer-reviewed by the operator among the plurality of contents. and, of the words related to the plurality of contents to which the evaluation values are given, positive words more strongly related to the contents to which the positive evaluation values are given, and to the contents to which the negative evaluation values are given. and identifying the plurality of words so that they can be presented separately from more strongly associated negative words.

Further, receiving the keyword information in the technology disclosed includes receiving at least one modification of a modification of the specified plurality of keywords and a modification of the corresponding weights, and determining the priority. This may include changing the priority to present the change in priority corresponding to content given the rating value in response to receiving the modification.

Further, changing the priority in the technology disclosed herein may include associating the evaluation value with the priority so as to allow an operator to identify the evaluation value.

Further, the technology disclosed may be a program that causes a computer to execute the above method.

According to the disclosed technique, when an operator is provided with a set of contents obtained by searching for contents including text, the operator can more efficiently grasp each content included in the set of contents. , the operator's labor can be reduced.

FIGS. 1A and 1B show various collections of content obtained by searching or the like. FIG. 2A is an example of information on which positive keyword selection can be based. FIG. 2B shows a table in which the operator has specified positive keywords and corresponding weights (r). FIG. 3A is an example of a graph that can be used as a basis for selecting negative keywords. FIG. 3B shows a table in which the operator has specified negative keywords and corresponding weights (r). FIG. 4A is a table in which the sum of the products of the number of keyword appearances of each content included in the set V and the weight corresponding to the keyword is arranged in descending order. FIG. 4B shows the results of peer review of all contents included in set V. FIG. 5A and 5B show an example of a user interface that assists the operator in changing the selection of keywords and corresponding weight changes. FIG. 6A shows the processing flow of the embodiment. FIG. 6B is a flowchart showing detailed processing of step S102 for estimating a plurality of contents. FIG. 7 is a flow chart showing detailed processing of step S112 in which a plurality of extracted words are presented to the operator. FIG. 8A is a flow chart showing the detailed process of step S104 of receiving keyword information, including a plurality of keywords and corresponding weights (including positive or negative). FIG. 8B is a flowchart showing detailed processing of step S108 for determining the priority of presentation of a plurality of contents based on the sum corresponding to each content. FIG. 9 is a flowchart showing detailed processing of the processing in step S312 for changing the priority so as to present a change in priority corresponding to the content given the evaluation value in response to receiving the correction. be. FIG. 10 is a block diagram illustrating the functionality of the embodiment. FIG. 11 is a diagram showing each hardware configuration of the embodiment.

In particular, when searching for patent documents, it is necessary to devise a search formula so as not to omit related patent documents and not to include many unnecessary patents (noise documents). Therefore, search formulas are examined to obtain a set of patent documents. However, many patent documents (noise documents) that are unrelated to the research target exist in the set of patent documents acquired by the search formula with such a device.

In order to reduce this noise literature, it is necessary to strengthen the search refinement. However, there is a risk that important documents will be omitted from the search results by strengthening the narrowing down of the search. Conversely, if a search is performed so as to prevent the omission of important documents, the set of documents obtained as a result of the search will become large, and the operator will have to spend a lot of effort in browsing (peer review).

For example, when extracting related patents by peer review, it is common to focus on whether or not words related to the target technical field are included. For example, if it contains words unrelated to the research subject, it is often judged as noise literature.
Therefore, in many cases, appropriate search results can be obtained by appropriately selecting words and phrases related to the research object and words and phrases not related to the research object, creating a search formula, and performing a search.

However, it should be noted here that in a patent document composed of long sentences, descriptions other than the target technology of the patent document may be mixed in part of the sentence. For example, there are many examples where examples of experiments that did not go well (counterexamples) are described. Alternatively, terms such as high or low may be used to express the degree of technical performance. When a term characterizing performance is used as a search term, the document may become noise if the term expressing the degree of performance does not match the required technology.

Note that the search formula includes a NOT operation. Conventionally, by specifying irrelevant words and phrases and incorporating a NOT operation into the search formula, retrieval results are obtained excluding documents that include irrelevant words and phrases. However, with this method, if the above-mentioned counterexamples are included in the literature, there is a risk that important literature will be omitted from the search results against the intention of the searcher.

Therefore, the disclosed technique proposes giving priority to each content included in a set of content by using weights corresponding to keywords related and unrelated to the research target desired by the operator. are doing. By using the priority to adjust the order in which the content is presented to the operator or to display the content with the priority associated with it, it is possible to facilitate the use of the content by the operator.

Note that the keywords used in this embodiment are keywords that can be set separately from the keywords used in the search formula, and are not necessarily the same keywords. The set of contents targeted by the present embodiment may be obtained by searching using a search formula including keywords, or may be obtained by other means, such as a set of contents collected using AI. may be That is, the present embodiment does not depend on collection means for collection of targeted content.
In the disclosed embodiment, content means an expression including content expressed using language, such as text, image, video, and audio.

An operator who drafts a search formula has a certain level of knowledge and understanding of technical terms. Thus, the operator can specify keywords that are closely related to content that the operator deems important. In addition, it is also possible to specify keywords that are closely related to content (noisy literature) that the operator deems unimportant. In addition, it is considered possible to specify synonyms and synonyms for each keyword.

Alternatively, a person who has the technical level to create a search formula can often select (or specify) related keywords and unrelated keywords without reviewing the content. In addition, if the operator had reviewed the content of some of the search results, he would be better able to select (or specify) keywords related and unrelated to the content of interest.

Furthermore, if a list of words and phrases included in a set of search results (or peer-reviewed) content is presented to the operator in an easy-to-understand manner by text mining, statistical processing, etc., the operator can find relevant keywords and unrelated keywords. Keyword selection becomes easier. In this case, the operator cannot predict what keywords will be included in the target content. In addition, there may be a case where the target content contains many synonyms, synonyms, and/or different notation of the keyword used by the operator in the search formula or the like. Therefore, it is possible to improve the accuracy of this method by selecting keywords from a list of words included in the target.

In this specification, a keyword closely related to content (positive content) (documents targeted for search) that the operator determines to be important is referred to as a positive keyword. Keywords that are closely related to content (negative content) (documents that become noise) that are judged to be unimportant to the operator are called negative keywords.
<Description on collection of search results>
Figures 1A and 1B show various collections of content obtained by searching.

As shown in FIG. 1A, a set V of results narrowed down using a search formula or the like is obtained for a population U of contents. Usually, the operator judges whether each content is important or not by sequentially reviewing the content included in the set V narrowed down by a search or the like.

A positive set R1 in FIG. 1A indicates a set including contents judged to be important among the contents of the set V reviewed by the operator. The positive set R1 is defined as a positive set including content (positive content) given a positive evaluation as a result of operator review.

The negative set G1 indicates a set that includes content that is judged to be unimportant among the contents of the set V reviewed by the operator. The negative set G1 is defined as a negative set including content (negative content) given a negative evaluation as a result of operator review.

Other set T1 indicates a set of contents that are neither important nor unnecessary (or that review has not been sufficiently performed and the evaluation result is not sufficient) among the contents of set V that have been peer-reviewed by the operator. The other set T1 is defined as the other set that includes content that has been rated neither negative nor positive by the results of the operator's review.

Usually, the operator sequentially reviews the contents included in the narrowing result set V, reviews all the contents in the set V, and evaluates each of the contents positively, negatively, or otherwise. etc. will be given. Note that the viewpoint of evaluation may change depending on the purpose of the investigation conducted by the operator. The purpose of the investigation is wide-ranging, including acquisition of patent rights, acquisition of materials to invalidate other companies' patents, prevention of infringement, understanding of other companies' technologies, and acquisition of basic materials for research and development. Needless to say, the viewpoints for judging the degree of importance (priority) of the same content differ according to these research purposes.

FIG. 1B shows an example of the state when the operator has finished reviewing all of the content for the set V of search results. That is, set V includes positive set R0, negative set G0, and other sets T0.
To obtain each set shown in FIG. 1B, it may generally be desirable for the operator to review all of the contents of set V. FIG.
<Embodiment>

In the embodiments described below, at least the positive set R0 can be pre-estimated by giving higher priority to content that is presumed to be more important to the operator.

By using this embodiment, the operator can refer to the priority given to the content and preferentially browse (review) the content that is likely to be important to the operator in advance. The operator can appropriately process the contents belonging to the set V more easily and in a short period of time by handling the contents in descending order of priority.
FIG. 2A is an example of information on which positive keyword selection can be based.
In FIG. 2A, keywords are listed on the horizontal axis, and "keyword index 1" corresponding to each keyword is plotted on the vertical axis.

Referring to the graph of FIG. 2A, the operator can select at least one positive keyword. Examples of "keyword index 1" in FIG. 2A include the following.

- Example of keyword index 1 1 . 1. The number of occurrences of words included in some or all of the contents in the set V of search results; 3. Average number of word occurrences per content that was given a positive rating in peer review. [Average number of word occurrences per content rated positive in peer review] - [Average number of word occurrences per content rated negative in peer review]

In the above description, the "number of appearances" may be the number of appearances of a word in a part of the content instead of the number of appearances of the word in the entire content. For example, when a patent document is included in the content, the number of occurrences of words may be counted by limiting the scope of claims of the patent document.
FIG. 2B shows a table in which the operator has specified positive keywords and corresponding weights (r).

The weight (r) corresponding to a keyword is preferably set to a value greater than or equal to zero as the keyword is included in content that is more important (that is, positive) to the operator. Then, the weight (r) can be set to be less than or equal to zero and have a larger absolute value as the keyword is included in content that is not important to the operator (that is, is negative). desirable. How to use the weight (r) will be described later.

If the content has not yet been peer-reviewed, the operator may not have the graph in FIG. 2A. regardless), the operator may freely specify positive keywords and corresponding weights (r) based on operator knowledge and the like. Alternatively, the operator may be presented with a list of words included in some or all of the contents belonging to set V, and the operator may select positive keywords from the list. At least one positive keyword is desirably specified.
Note that if the weight (r) is specified as zero, it is the same as if the corresponding positive or negative keyword was not specified. Therefore, by designating the weight (r) as zero, it is possible to do the same thing as canceling the designation of a positive keyword or a negative keyword, which simplifies the operation of canceling the designated keyword.

The "synonyms/different notations" may be automatically specified by the computer by referring to a dictionary. Alternatively, by referring to a dictionary by a computer, candidates for "synonyms/different notations" may be presented to the operator, and the operator may be prompted to make a selection. Alternatively, it may be set by the operator. It is desirable that the words specified as "synonyms/different notations" be treated in the same way (as the same word) as the corresponding positive keywords.
FIG. 3A is an example of a graph that can be used as a basis for selecting negative keywords.
In FIG. 3A, “keyword index 2” is plotted on the vertical axis corresponding to each word.

Referring to the graph of FIG. 3A, the operator can select at least one negative keyword. Examples of "keyword index 2" in FIG. 3A include the following.
- Example of keyword index 2 4 . 5. The number of appearances of words included in some or all of the contents in the set V of search results; Average number of word occurrences per piece of content rated negative in peer review

6. [Average number of word occurrences per content rated negative in peer review] - [Average number of word occurrences per content rated positive in peer review]
FIG. 3B shows a table in which the operator has specified negative keywords and corresponding weights (r).

The weight (r) is desirably set to be negative and have a large absolute value as the content is less important (that is, negative) to the operator. How to use the weight (r) will be described later.

It should be noted that the operator may not have the graph in FIG. regardless), the operator may freely specify negative keywords and corresponding weights (r) based on operator knowledge and the like. Alternatively, the operator may be presented with a list of words included in some or all of the contents belonging to set V, and the operator may select negative keywords from the list. At least one negative keyword is desirably specified. If the weight (r) is specified as zero, it is the same as if the corresponding negative keyword was not specified.

The "synonyms/different notations" may be automatically specified by the computer by referring to a dictionary. Alternatively, by referring to a dictionary by a computer, candidates for "synonyms/different notations" may be presented to the operator, and the operator may be prompted to make a selection. Alternatively, it may be set by the operator. It is desirable that the words specified as "synonyms/different notations" be treated in the same way (as the same word) as the corresponding negative keywords.
Note that the above positive evaluation and negative evaluation are examples of evaluation values.
FIG. 4 is a table in which the totals (Total(m)) corresponding to each content included in the set V of search results are arranged in descending order.

ただし、
ｍ：検索結果の集合Vに含まれる各々のコンテンツを識別する番号（１から始まる連続する整数のいずれか）
u：ポジティブキーワードの数＋ネガティブキーワードの数、すなわちキーワードの総数
c(m)_n：コンテンツｍにおける、 n番目のポジティブキーワード又はネガティブキーワード（キーワード）の出現回数
r_n：n番目のキーワードに対応する重みr
For example, in the content m belonging to the set V of the search results, the sum of the products of the number of appearances of the keyword and the weight corresponding to the keyword for all the keywords, that is, Total(m) is defined as follows.

however,
m: A number that identifies each content included in the search result set V (one of consecutive integers starting from 1)
u: number of positive keywords + number of negative keywords, i.e. total number of keywords
c(m) _n : Number of appearances of the nth positive or negative keyword (keyword) in content m
r _n : weight r corresponding to the nth keyword

The above total, that is, the sum Total(m) of the product of the number of appearances of the keyword of the content m included in the search result set V and the weight corresponding to the keyword is an example of the priority of the content m.

The weight corresponding to the positive keyword is preferably a numerical value greater than or equal to zero, and the weight corresponding to the negative keyword is preferably a numerical value less than or equal to zero.

It can be estimated that the greater the number of appearances of a certain positive keyword in a certain content, the higher the possibility that the content is more closely related to the details of the technique desired by the operator. In addition, the higher the weight of the positive keyword, the higher the possibility that the content containing the positive keyword is the content closer to the technology desired by the operator.

It can be inferred that the greater the number of appearances of a certain negative keyword in a certain content, the higher the possibility that the content is more distantly related to the details of the technology desired by the operator. In addition, the greater the absolute value of the negative keyword weight (value less than or equal to zero), the higher the possibility that the content containing the negative keyword is content that is farther from the content of the technology desired by the operator.

Therefore, the product of the number of appearances of a certain keyword (positive keyword or negative keyword) in certain content and the weight corresponding to that keyword is an index of how important that content is to the operator. Then, for all keywords (positive keywords or negative keywords) contained in a certain content, the total sum of the products is an index (priority degrees).
Therefore, it can be estimated that the larger the total value (priority) corresponding to the content is, the closer the content is to the technology desired by the operator.
FIG. 4A is a table in which the sum of the products of the number of keyword appearances of each content included in the set V and the weight corresponding to the keyword is arranged in descending order.

In the table 400 of FIG. 4A, it can be estimated that the higher the content is positioned in the table, the higher the priority of the content. Conversely, it can be estimated that the lower the content is located in the table, the lower the priority of the content.

For example, the content with the content number 45 positioned at the top of the table 400 in FIG. 4A is the content with the largest total Total(w) value. This content means that the sum of the products of the number of occurrences of the keywords (positive keywords and negative keywords) and the corresponding weights is the largest. In the table 400 of FIG. 4A, the higher the content ranks, the more weighted positive keywords are included in the content, and the content is determined to have a higher priority.

FIG. 4B shows the results of peer review of all the contents included in set V. As shown in FIG. The horizontal axis indicates the order of priority. The vertical axis shows the ratio of content judged to be important among m pieces of content with priority (No.) from 1 (No.=1) to priority m (content judged to be important). It is a graph showing the content rate).

For example, in graph 410, 43 contents from priority 1 to priority 43 shown on the horizontal axis, and 90% of the contents belonging to set V judged to be important. means that it contains By giving priority to contents using this embodiment and reviewing 43 high-priority contents out of 200 contents belonging to set V of search results, 90 contents in set V are reviewed. This means that 100% of important (high priority) content can be found through peer review.

Furthermore, by reviewing 100 pieces of content in descending order of priority, the operator can discover 100% of the content that belongs to set V and is judged to be important. In this manner, the present embodiment can provide the operator with a priority of content so as to efficiently review content from among the 200 content contained in the set V of search results.
FIG. 5 shows an example of a user interface that assists the operator in changing the selection of keywords and changing the corresponding weights.

In FIG. 5, it is assumed that 17 contents out of 200 contents included in set V have been peer-reviewed. For example, 6 of the contents are closely related to the operator's purpose and judged to be important (positive set R1 in FIG. 1), and 8 contents are not relevant to the operator's purpose and judged to be unnecessary. (negative set G1 in FIG. 1) and the content of 3 was determined to be in the Other category (Other set T1 in FIG. 1).

A peer-reviewed content table 520 in FIG. 5A is a table in which the Total(m) of the peer-reviewed content is arranged in descending order. Column i displays black, white, and hatching corresponding to the set (R1, G1, or T1) to which the content shown in FIG. 1 belongs, respectively. While referring to the identification in this column i, the operator modifies the keywords or the corresponding weights shown in table 520 so that most of the contents belonging to positive set R1 are positioned at the top of table 520 as much as possible. can do. Among the peer-reviewed documents, content belonging to the positive set R1 is placed at the top of the table 520, so that the value of Total(m) can be made larger for positive content than for non-peer-reviewed content. can increase This can increase the likelihood that positive non-peer-reviewed content can be given higher priority.

Table 500 of FIG. 5A displays previously specified positive and negative keywords and corresponding weights 502 . The operator can add new keywords to table 520 or delete keywords. In addition, a user interface may be provided that allows the operator to easily change the weight by sliding the mark 506 on the slide bar 504 left and right with a mouse or the like. Alternatively, the operator may be allowed to change the weight by key input or the like.
Note that setting the weight to zero has the same effect as deleting a keyword (ie, excluding it from consideration).
FIG. 5B is a diagram showing a state in which the order of contents is changed in accordance with the change in the display of Total(m) and the i column when the keyword weight is changed.

The peer-reviewed content table 501 of FIG. 5B moves the position of the mark 506 of the keyword "heater" in the peer-reviewed content table 500 of FIG. 1” is shown. In response to this weight change, the order in which the contents are arranged in the table 521 of FIG. 5B is shown to be changed according to the value of Total(m).

As compared to the peer-reviewed content table 520 of FIG. 5A, in the peer-reviewed content table 521 of FIG. . It can also be seen that the lowest content (m=112) in the positive set in table 520 of FIG. 5A has risen from 11th place in table 520 of FIG. 5A to 7th place in table 521 of FIG. 5B.

From the above, it is possible for the operator to easily recognize that the pattern of association of weights corresponding to multiple keywords in table 501 in FIG. 5B is more preferable than the pattern in table 500 in FIG. 5A.
Deletion and addition of keywords will not be described with reference to drawings, but those skilled in the art will understand that the arrangement of content changes according to keyword correction.

The operator attempts to modify the keywords or the corresponding weights accordingly so that most of the content belonging to the positive set R1 of peer-reviewed content is positioned as high as possible in the table 521 of FIG. 5B. can be

The user interface in FIGS. 5A and 5B is an example and is not limited to this example. In FIG. 5, only peer-reviewed content is included in table 520 of FIG. 5A and table 521 of FIG. 5B, but all content belonging to set V may be included. In this case as well, the operator may correct the keywords or the weights so that the contents belonging to the peer-reviewed positive set R1 are positioned as high in the table as possible. Alternatively, the operator may correct the keywords or the weights so that the lowest content among the content belonging to the peer-reviewed positive set R1 is ranked as high as possible.

Then, a pattern of weights corresponding to keywords recognized as being preferred by the operator is determined. The determined pattern is used to provide the operator with the sum of the contents belonging to set V (ie the priority). By reviewing the content in order of priority, the operator can preferentially review content that is presumed to be important to the operator.
FIG. 6A shows the processing flow of the embodiment. The processing flow shown in FIG. 6A will be described below.

[Step S102] A plurality of contents (that is, contents belonging to the set V) are identified. For example, an operator obtains a set V of search results from a content population U using a search formula, thereby identifying a plurality of target contents.
[Step S104] Keyword information is received, including a plurality of keywords and corresponding weights (including positive or negative).

[Step S106] The product of the number of occurrences of the keyword and the corresponding weight is calculated for each of the plurality of contents, and the sum of the products is derived for the plurality of keywords.

[Step S108] The priority of presentation of a plurality of contents is determined based on the sum corresponding to each content. Note that this total (priority) is not limited to the priority of the order in which content is presented to the operator. For example, the total (priority) may be displayed and shown to the operator along with the display of the content so that the operator can recognize the estimated importance of the content being displayed.

The operator can efficiently process the contents belonging to the set V based on the sum (priority) corresponding to each of the contents obtained by the above processing.
FIG. 6B is a flowchart showing detailed processing of step S102 for estimating a plurality of contents. The processing will be described below.

[Step S112] A plurality of words related to a plurality of contents are extracted. The plurality of contents may be all the contents belonging to the set V, or may be a part of the contents. Alternatively, words may be extracted from a portion of the content (for example, the scope of claims in the case of patent documents).

[Step S114] The plurality of extracted words are presented to the operator. Positive or negative keywords may be identified based on the presented words. Alternatively, a positive or negative keyword determined by the operator himself may be identified.
Presenting the words to the operator facilitates the operator's identification of positive or negative keywords.
FIG. 7 is a flow chart showing detailed processing of step S112 for extracting a plurality of words related to a plurality of contents. The processing will be described below.

[Step S202] A positive or negative evaluation value given by the operator to each of the plurality of contents peer-reviewed by the operator is received. Through this process, the positive set R1 and the negative set G1 shown in FIG. 1 are identified. Other sets T1 may be specified. Note that the other set T1 does not necessarily have to be specified.

[Step S204] Among words related to a plurality of content given evaluation values, positive words more strongly related to content given a positive evaluation value and content given a negative evaluation value , multiple words are identified so that they can be presented separately from the more strongly associated negative words.

As already described in FIGS. 2 and 3, this process facilitates the operator to determine positive or negative keywords or their corresponding weights.

FIG. 8A is a flow chart showing the detailed process of step S104 of receiving keyword information, including a plurality of keywords and corresponding weights (including positive or negative). The processing will be described below.

[Step S302] Modifications of at least one of a plurality of designated keywords and corresponding weight modifications are received. This process yields more favorable keywords (positive or negative keywords) or corresponding weights.

FIG. 8B is a flowchart showing detailed processing of step S108 for determining the priority of presentation of a plurality of contents based on the sum corresponding to each content. The processing will be described below.
[Step S312] In response to receiving the modification, the priority is changed to present a change in priority corresponding to the content given the rating value.
Through this process, each content belonging to the set V is assigned a priority desired by the operator.
The operator can use this priority to efficiently perform content review and the like.

FIG. 9 is a flowchart showing detailed processing of the processing in step S312 for changing the priority so as to present a change in priority corresponding to the content given the evaluation value in response to receiving the correction. be. The processing will be described below.
[Step S402] The priority is associated with the evaluation value so that the operator can identify the evaluation value. As described with reference to FIGS. 5A and 5B, by performing this processing, the operator can add keywords (positive keywords or negative keywords) can be determined with modifications or corresponding weights.

FIG. 10 is a block diagram illustrating the functionality of the embodiment. This block diagram will be described below.
The content specifying unit 1002, for example, specifies various pieces of information about the content from the search results.

The word extraction unit 1004 receives, for example, a positive set R1, a negative set G1, or another set T1, extracts words existing in these sets, and presents them to the operator. The word extraction unit 1004 may extract words from all the contents belonging to the set V or from some of the contents. Note that the word extraction unit 1004 may not function in some cases. In this case, the operator may operate the keyword specifying unit 1006 and the weight determining unit 1008 to specify the keyword and its weight.

A keyword (positive keyword or negative keyword) is identified by the keyword identification unit 1006 . Keywords may be selected by the operator from a presented list of words. Alternatively, a keyword specified by the operator himself may be used.
The weight determination unit 1008 can determine weights corresponding to keywords based on instructions (or correction instructions) from the operator.
The dictionary storage unit 1010 is used to extract synonyms, synonyms, and/or different notation keywords corresponding to the keyword.
The content priority determining unit 1012 calculates the total (priority) corresponding to the content as described above.
The calculated priority is used by the operator to efficiently process the content.

FIG. 11 is a diagram showing each hardware configuration of the embodiment. The hardware configuration includes a CPU 3001 , a ROM 3002 in which programs, databases and/or data of this embodiment can be stored, a RAM 3003 , a network interface 3005 , an input interface 3006 , a display interface 3007 and an external memory interface 3008 . These pieces of hardware are interconnected by a bus 3004 .

Network interface 3005 is connected to network 3015 . The network 3015 includes a wired LAN, a wireless LAN, the Internet, a telephone network, and the like. An input unit 3016 is connected to the input interface 3006 . A display unit 3017 is connected to the display interface 3007 . A storage medium 3018 is connected to the external memory interface 3008 . The storage medium 3018 may be RAM, ROM, CD-ROM, DVD-ROM, hard disk, memory card, USB memory, or the like.
Programs and methods that implement the above-described embodiments can be executed by a computer having the hardware configuration shown in FIG.

Each of the embodiments described above is not exclusive, and a part of one embodiment may be incorporated into another embodiment, or a part of one embodiment may be substituted for another embodiment. can be

Additionally, each flow in the illustrated flow charts can be interchanged in order as long as there is no conflict. Also, as long as there is no contradiction, one illustrated flow can be executed multiple times at different timings. Multiple steps may be performed simultaneously. Each step may be implemented by executing a program stored in memory (non-transitory memory).

Also, some programs of the disclosed embodiments can be implemented in general-purpose programs such as operating systems or hardware. In addition, the disclosed programs may be distributed and executed on multiple pieces of hardware.

A program that implements the above-described embodiments can be executed by a computer having the hardware configuration shown in FIG. Also, the program of the embodiment may be implemented as a method to be executed by a computer.

It goes without saying that the above embodiments do not limit the invention described in the claims, but are treated as examples. In addition, the text and voice that can be included in the content targeted by the disclosed technique and the claimed invention are not limited to a specific language, and can be expressed in any language. Alternatively, multiple languages may be mixed.

1002 Content identification unit 1004 Word extraction unit 1006 Keyword identification unit 1008 Determination unit 1010 Dictionary storage unit 1012 Content priority determination unit

Claims (9)

A content processing method in which a computer executes content processing for determining the priority of presentation of each of a plurality of content,
identifying the plurality of pieces of content, including receiving a rating value given by the operator to each of the plurality of content peer-reviewed by the operator among the plurality of pieces of content;
Receiving keyword information including a plurality of keywords designated by the operator and weights corresponding to each of the plurality of keywords, and modifying the weights corresponding to the designated plurality of keywords. to receive, including to receive;
For each of the plurality of contents, adding the product of the number of appearances of each of the plurality of keywords and the weight corresponding to the keyword for each of the plurality of keywords to obtain a total corresponds to each of the plurality of contents. and deriving the sum of
determining a priority for presentation of each of the plurality of pieces of content based on the sum corresponding to each of the plurality of pieces of content, and in response to receiving a modification of the weights ; changing the presentation priority based on the sum derived using the modification, and recognizing the change in the presentation priority corresponding to the content given the rating value in association with the rating value by the operator; determining, including displaying on a display to cause
A content processing method comprising: The evaluation value includes a positive evaluation value or a negative evaluation value,
The content processing method according to claim 1. receiving the keyword information further includes receiving modifications of the plurality of keywords;
The content processing method according to claim 1 or 2. Identifying the plurality of contents includes:
extracting a plurality of words associated with the plurality of content from the plurality of content ;
presenting the plurality of words to an operator such that the operator identifies the plurality of keywords from the plurality of words based on the plurality of words;
4. The content processing method according to any one of claims 1 to 3, comprising: In the extracting
extracting the plurality of words from a predetermined portion of the plurality of content ;
The content processing method according to claim 4. The extracting includes
Among the words related to the plurality of contents given the evaluation values, positive words more strongly related to the contents given the positive evaluation values, and more positive words to the contents given the negative evaluation values. identifying the plurality of words such that they are presented separately from strongly associated negative words;
6. The content processing method according to claim 4 or 5, comprising: the weight corresponding to each of the plurality of keywords is a positive value (including zero) or a negative value (including zero);
7. A content processing method according to any one of claims 1 to 6. the content includes at least one of text, images, and audio;
A content processing method according to any one of claims 1 to 7. A program that causes a computer to execute the method according to any one of claims 1 to 8.

Images